ABSTRACT
This chapter provides examples of how Francis Galton’s linear correlation coefficient was at first ignored by the scientific community. However, within a few years, it began to appear in the publications of a few biologists, economists, and mathematicians. Those papers proposed names and formulas for Galton’s new coefficient, given that Galton had not done so. Some of those formulas (e.g., by Walter Weldon and Francis Edgeworth) were so very crude that their output could sometimes be worse than a pure guess. However, the formula devised by Karl Pearson was so very exact that it became the most-used definition of correlation, and it still is today; he also devised a formulaic method to calculate a coefficient for curvilinear correlation (which he called the “correlation ratio”). One scientist devised a new graphical way to estimate the correlation coefficient, which involved plotting two different regression lines on the same chart and then measuring the angle that they made with each other; this too provided a crude approximation to the true value. The concept of mathematical correlation was used to support the use by police of the new science of finger prints (which had been recently developed by Galton), and was used to denigrate the use of a rival system of biometrics (developed by Alphonse Bertillon) that was at that time well-established world-wide.
